In his research, Germain Gauthier develops and applies machine learning models to study topics at the intersection of political science and economics. He is particularly invested in natural language processing techniques (“text as data”) for social scientists. In his methodological work, he has proposed novel tools to mine political and economic narratives from large text corpora, as well as neural topic models for massive collections of documents with metadata.
In his applied work, Germain Gauthier combines natural language processing techniques with (quasi-)experimental evidence to study a broad set of topics in political economy. Recently, he has written about the consequences of social media on political attitudes, social movements, and gender norms, and studied the impact of Large Language Models on gender stereotypes in the labor market. At CESifo, Mr. Gauthier will present these ongoing projects and discuss economists’ needs and expectations regarding machine learning techniques for Big Data.